GPU Acceleration in LAMMPS
GPU Acceleration in LAMMPS
GPU Acceleration in LAMMPS
- No tags were found...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
M<strong>in</strong>imiz<strong>in</strong>g the Code Burden • Advantages • Allow a legacy code to take advantage of accelerators without rewri1ng the en1re th<strong>in</strong>g • In some cases, there is no advantage to por1ng • Mul1-‐Core CPUs can be compe11ve with <strong>GPU</strong> accelerators for some rou1nes (latency-‐bound, thread divergence, etc.), especially at lower par1cle counts • If concurrent CPU/<strong>GPU</strong> calcula1ons are used effec1vely, there can be no advantage to por1ng certa<strong>in</strong> rou1nes • For some simula1ons, the upper bound for performance improvements due to por1ng addi1onal rou1nes and remov<strong>in</strong>g data transfer is < 5%.